Citi’s Robotics & Bodily AI Management Convention wrapped up on Tuesday. The annual Citi Analysis occasion brings collectively robotics founders, traders, operators, and trade executives to assess the state of “bodily AI.”
Analyst Heath Terry summarized the important thing takeaways Wednesday morning, portray an image of the robotics trade shifting from proof of idea to business deployment, whereas warning that scaling robots stays difficult.
“Labor shortages, reshoring, and favorable regulatory tailwinds are accelerating enterprise demand, whereas knowledge shortage, expertise constraints, battery limitations, and excessive deployment prices stay key friction factors,” Terry defined to shoppers.
Citi stated the winners in bodily AI will possible be companies that personal proprietary real-world knowledge, resolve particular labor bottlenecks and use Robotics-as-a-Service fashions to scale back upfront prices for purchasers.
Terry highlighted automation-exposed industrial names together with Rockwell Automation, Emerson Electrical, Honeywell, Symbotic, Ralliant and Belden as potential beneficiaries.
Humanoids are attracting vital investor curiosity. Final month, we detailed how readers can make investments forward of a serious ramp in humanoid manufacturing anticipated within the coming quarters. Read the report.
By way of Deutsche Bank:

During the last two years, about $20 billion has been invested in bodily AI, with functions spanning warehouses, logistics, trucking, development, aviation, and protection.
Final week, carmaker BMW revealed {that a} new upgraded humanoid is strolling its manufacturing facility flooring on the Spartanburg plant in South Carolina.

Listed here are Citi’s prime takeaways from the bodily AI convention:
Key AI Takeaways
Bodily AI is transitioning from proof-of-concept to business deployment, however the path to scale is extra operationally intensive than the digital AI analogy suggests. Not like giant language fashions, the place a base mannequin carries numerous the worth, bodily AI locations the premium on proprietary, task-specific knowledge collected in real-world environments, purpose-built {hardware}, and security certification.
Throughout classes, members constantly recognized knowledge shortage because the binding constraint, with Instawork noting that even tens of tens of millions of hours of knowledge being collected in 2026 possible represents solely foundation factors, not share factors, of what’s in the end wanted to attain high-level robotic efficiency. Energy, battery longevity, and chip structure are additionally rising as crucial bottlenecks, with panelists noting that present semiconductor platforms had been designed for datacenter workloads, not real-time edge inference on cell platforms.
Essentially the most commercially superior corporations, whether or not in humanoids, warehouse AMRs, autonomous trucking, or development, shared a standard profile: they began with a particular, high-pain labor drawback, adopted a Robotics-as-a-Service mannequin to decrease buyer adoption limitations, and prioritized security and reliability above mannequin sophistication.g vital funding enthusiasm, however near-term ROI is being pushed by purpose-built AMRs and specialised methods from corporations like Locus Robotics and Dexterity. The convention strengthened our view that bodily AI is a decade- lengthy buildout, with sturdy worth accruing to corporations that personal the information flywheel, resolve actual deployment issues, and meet the very best security requirements.
Key Industrials Takeaways
After attending Citi’s Robotics & Bodily AI Management Convention, we got here away satisfied that automation, robotics, and bodily AI proceed to make gradual progress towards broader commercialization, creating what we view as a sturdy long-term development tailwind for our automation-exposed industrial corporations. Our most well-liked methods to realize publicity to industrial automation are via Purchase-rated ROK, EMR, and HON (pure-play automation suppliers), SYM (warehouse automation), RAL (sensors and T&M), and BDC (industrial networking), as we view these corporations as effectively positioned to learn from rising investments in automation (together with tools, software program, and AI). Key drivers of automation adoption stay a constrained labor market in addition to accelerating home manufacturing exercise and capability expansions, with automation supportive of upper throughput, elevated uptime, and improved operational effectivity/accuracy that seems supportive of wholesome ROIs.
Whereas logistics, warehousing, and autos seem like necessary finish markets driving automation adoption (increased quantity, repetitive duties), a number of panelists highlighted AI’s potential to unlock new expertise/capabilities for robotics and automation, which may, over time, develop addressable markets for automation options (new finish markets and new use instances), which we view as a broad constructive. Advances in AI/LLMs together with rising availability of knowledge (real-world in addition to simulated) are resulting in extra refined know-how via, as an illustration extra built-in {hardware} and software program and extra energetic utilization driving more and more succesful deployments (as AI-enabled methods develop “smarter”), which we predict may additional assist accelerating automation adoption over time. In some situations, entry to high-quality knowledge stays a bottleneck (the place corporations are combining simulations with restricted real-world knowledge), however we additionally view this as a chance and aggressive moat for our corporations, given their giant put in base and talent to leverage/mine knowledge to additional drive autonomy and operational effectivity.
Choose corporations additionally highlighted robotics-as-a-service (RaaS) enterprise fashions that doubtlessly decrease the limitations to adoption for small and medium sized enterprises given decrease or minimal upfront capital prices, which we view as supportive of our constructive view on SYM’s warehouse-as-a-service providing (GreenBox/Exol), which we predict may assist drive rising adoption of SYM’ warehouse automation options amongst a broader vary of consumers.
The most necessary takeaway is that Bodily AI is rising and deployments might be ramped up as Robotics-as-a-Service fashions scale back upfront buyer prices. Bodily AI could ultimately profit from scaling legal guidelines, however the path might be a lot slower and extra operationally intensive than the chatbot growth.
By Zerohedge









